Mobile Menu

Discovery of Candidate Antibody-Drug Conjugate Targets

A recent study used the Human Protein Atlas as a unique big data resource to identify and prioritise candidate antibody-drug conjugate targets with translational potential across common types of cancer.

Antibody-drug conjugates

There is a lot of interest surrounding research on antibody-drug conjugates as a promising targeted therapy for cancer. However, despite recent advances in the field, only a few antibody-drug conjugates have been approved by the FDA. This is largely due to the lack of enough tumour response data or excessive normal tissue toxicity observed in clinical trials. Therefore, the expansion of effective and non-toxic antibody-drug conjugates is still a challenge for drug developers, and the discovery of novel antibody-drug conjugate targets is of high interest.

Selecting molecular targets for antibody-drug conjugates

The selection of appropriate target antigens is the first critical set in the development of safe and effective antibody-drug conjugates. At present, there are only a few large-scale studies that have identified or prioritised antibody-drug conjugate targets. As far as the researchers are aware, no big data research based on the protein-level evidence exists for identification and prioritisation of candidate antibody-drug conjugates across a wide range of tumour types.

The Human Protein Atlas (HPA) is a large-scale antibody-based proteomic resource, which provides a unique opportunity to perform systematic discovery and validation of targets for different tumour types at the protein level. The HPA combines the antibody-based approach with transcriptomic data for an overview of global expression profiles. In this study, the researchers aimed to identify and prioritise molecular targets for antibody-drug conjugates per common tumour types by mining the HPA database.

Discovery of molecular targets for antibody-drug targets via screening

To systematically identify and prioritise candidate antibody-drug conjugate targets across 20 different tumour types, the researcher applied the following screening approach:

The data from 19,670 genes encoding human proteins, and the expression data for 5,520 membrane protein-coding genes was downloaded from the HPA website. In total, 2,131 genes without the protein level evidence were excluded from the study, and the remaining 3,389 genes were monitored for their protein expression in the critical normal tissues. In addition to that, the protein-coding genes that showed high protein expression levels in one or more critical normal tissue (including lung, gastrointestinal tract, liver, kidney, heart muscle, skin, and bone marrow) were excluded. The remaining 1,654 genes were retained in the data mining process. Following that, the protein expression levels for the remaining genes were monitored across 20 tumour types based on data extracted from the HPA database. After calculating a quasi-score (ranging from 0 to 300) for each tumour type, 745 genes with a score greater than 150 for at least one tumour type were included in the next step. The quasi score acts a proxy for protein expression. In order to discriminate the target antigens localized on the cell surface from non-surface membrane proteins, the researchers used data containing predicted sets of human surfaceome. Consequently, 332 potential target genes encoding surface proteins were extracted and included in an HPA-based three-step validation process.

Identified molecular targets for antibody-drug conjugates

Following the nonexperimental validation process, 23 candidate ADC targets were identified and prioritised across 20 tumour types. Interestingly, systematic analysis identified 3 antibody-drug conjugates that targeted the researchers identified targets, which have already been approved by the FDA for use in solid tumours and lymphoma. Moreover, four more drugs targeting the validated targets have also entered clinical trials. These seven clinically relevant antibody-drug conjugate targets serve as verification of the researcher’s discovery approach and suggests that their list may also contain novel candidate targets with a high translational potential. Therefore, this study may help to shorten the therapeutic road from the laboratory to the clinic for antibody-drug conjugates.


The results of this study showed that mining the HPA database has the potential to identify and prioritise molecular targets for antibody drug conjugates with translational potential across different tumour types. Therefore, this study could aid in the selection of targets, which should be taken for further investigation and consequently lead to the development of new clinically effective and safe antibody-drug conjugates in the future.

Image credit: pikisuperstar – FreePik

More on these topics

Big Data / Cancer / Target Selection

Share this article